techscript

Compiler Architecture

This document describes the compilation pipeline of the TechScript 2.0 compiler.

---

🗺️ Pipeline Stages

graph TD
    A[Source File] --> B[Lexer]
    B --> C[Parser]
    C --> D[Abstract Syntax Tree]
    D --> E[Semantic Analysis]
    E --> F[IR Generation]
    F --> G[Code Generator]

1. Lexer (compiler/lexer)

Scans the source character stream and generates tokens using the logos Rust library. Whitespace is ignored except when separating keywords or within string literals.

2. Parser (compiler/parser)

Implements a top-down operator precedence (Pratt Parser) for parsing expressions cleanly. This avoids deeply nested recursive grammar structures and cleanly parses operator precedence.

3. Abstract Syntax Tree (compiler/ast)

The parsing phase outputs an AST containing nodes representing program statements, expressions, loops, and definitions. Every node holds source code locations (Span) for error attribution.

4. Semantic Analyzer (compiler/semantic)

Validates program semantics prior to runtime:

• Scope Resolution: Ensures variables are defined in the current block or enclosing parents.
• Constant Verification: Fails compilation if code attempts to reassign a constant.
• OOP checks: Verifies subclasses inherit from valid models, and trait contracts are fully implemented.

5. IR & Code Generation (compiler/ir, compiler/bytecode, compiler/llvm_backend)

Translates optimized AST structures into:

• Bytecode Instructions: 8-bit opcodes for the Virtual Machine (VM).
• LLVM IR: Compiled into platform-specific machine code.

This site is open source. Improve this page.